TAILIEUCHUNG - Báo cáo khoa học: "Bayesian Learning of a Tree Substitution Grammar"

Tree substitution grammars (TSGs) offer many advantages over context-free grammars (CFGs), but are hard to learn. Past approaches have resorted to heuristics. In this paper, we learn a TSG using Gibbs sampling with a nonparametric prior to control subtree size. The learned grammars perform significantly better than heuristically extracted ones on parsing accuracy. | Bayesian Learning of a Tree Substitution Grammar Matt Post and Daniel Gildea Department of Computer Science University of Rochester Rochester NY 14627 Abstract Tree substitution grammars TSGs offer many advantages over context-free grammars CFGs but are hard to learn. Past approaches have resorted to heuristics. In this paper we learn a TSG using Gibbs sampling with a nonparametric prior to control subtree size. The learned grammars perform significantly better than heuristically extracted ones on parsing accuracy. 1 Introduction Tree substition grammars TSGs have potential advantages over regular context-free grammars CFGs but there is no obvious way to learn these grammars. In particular learning procedures are not able to take direct advantage of manually annotated corpora like the Penn Treebank which are not marked for derivations and thus assume a standard CFG. Since different TSG derivations can produce the same parse tree learning procedures must guess the derivations the number of which is exponential in the tree size. This compels heuristic methods of subtree extraction or maximum likelihood estimators which tend to extract large subtrees that overfit the training data. These problems are common in natural language processing tasks that search for a hidden segmentation. Recently many groups have had success using Gibbs sampling to address the complexity issue and nonparametric priors to address the overfitting problem DeNero et al. 2008 Goldwater et al. 2009 . In this paper we apply these techniques to learn a tree substitution grammar evaluate it on the Wall Street Journal parsing task and compare it to previous work. 2 Model Tree substitution grammars TSGs extend CFGs and their probabilistic counterparts which concern us here by allowing nonterminals to be rewritten as subtrees of arbitrary size. Although nonterminal rewrites are still context-free in practice TSGs can loosen the independence assumptions of CFGs because larger rules capture more .

• Báo cáo khoa học
• Bayesian Learning
• Tree Substitution Grammar
• Matt Post and Daniel Gildea
• báo cáo khoa học
• báo cáo ngôn ngữ
• xử lý ngôn ngữ tự nhiên
• Journal of Science and Technology
• Learning optimal threshold for bayesian posterior probabilities
• Mitigate the class imbalance problem
• Bayesian posterior probabilities
• The UCI repository show
• Significantly better in certain cases
• Machine Learning Introduction
• Machine Learning Work
• Types of Machine Learning
• Supervised Learning
• Forms of Supervised Learning
• Bayesian Learning of Non compositional Phrases
• Synchronous Parsing
• Hao Zhang
• A Dynamic Bayesian Framework
• Model Context and Memory in Edit Distance Learning
• An Application to Pronunciation Classification
• mô hình ngôn ngữ
• Learning Stochastic OT Grammars A Bayesian approach
• Data Augmentation
• Gibbs Sampling
• Tạo chí tin học
• Điều khiển học
• A bayesian belief network
• Generate learning path
• Adaptive hypermedia system
• Hệ thống học thích nghi
• Deriving prior distributions
• Bayesian models used
• Achieve adaptive e learning
• Group profile and performance data
• Prior distribution
• BMC Bioinformatics
• Genomic transcriptional networks
• Bayesian networks
• Hybrid structure learning algorithm
• System biology
• Machine learning
• Neuron network
• Deep learning probabilistic models
• Specialized mini projects
• Bayesian neural networks
• Statistical Inference
• Bayesian Network
• Learning Paradigm
• Equivalence Classes
• EM Algorithm
• Bayesian Information
• Complex Bayesian
• Introduction to machine learning
• Bayesian decision theory
• Parametric methods
• Multivariate methods
• Artificial Intelligence
• Lecture Artificial Intelligence
• Trí tuệ nhân tạo
• Statistical learning
• Bayes net learning
• Lecture note Artificial Intelligence
• Bài giảng trí tuệ nhân tạo
• Khoa học máy tính
• Nucleotide Substitution
• Phylogenetic Tree
• Read Between the Lines
• Bayesian Logic Programs
• Long Papers
• ngôn ngữ tự nhiên
• BMC Cancer
• Multimarker models
• Bayesian statistical learning method
• Bayesian regression
• Determination coefficient
• Bladder cancer outcome
• Genome wide common SNP
• Optimal Bayesian classification
• Prior construction
• Biological pathways
• Probabilistic Boolean networks
• Gene regulatory network
• High order dynamic Bayesian Network
• Hidden common cause
• Causality inference
• Ebook Pattern Recognition and Machine Learning
• Pattern recognition
• Graphical models
• K means clustering
• Bayesian linear regression
• Principal component analysis
• Multiple Sentence
• Information Retrieval
• Prosody Learning
• Language Learning
• Quality Scores
• Bayesian Methods
• Testing effect
• Production effect
• Nonword learning
• Week following training
• Bayesian analysis
• Probabilistic assessments
• Support vector machine
• Biomedical data sets
• Báo cáo sinh học hay
• cách trình bày báo cáo
• báo cáo sinh học
• công trình nghiên cứu sinh học
• tài liệu về sinh học
• Bayesian approach
• etting systems
• random walks
• construction of fiber optics
• spread spectrum theory
• e learning communication
• telecommunication network design